In order to dissect the biochemical steps involved in genetic recombination we have chosen to focus on a key early step: strand exchange between homologous parental DNAs. The product of this strand exchange reaction,is a joint molecule composed of a single-strand circle joined to one end of a linear duplex. Three proteins responsible for this step have been purified: uvsx from phage T4; RecA from E. coli; and recl from U. maydis. Over the last few years we have reported the partial purification and Characterization of similar strand-exchange proteins or recombinases from nuclear extracts of human cells and tissues and embryos of D. melanogaster have shown that the structure of the protein-free intermediate in strand-- exchange is most likely that of a three-stranded DNA. Recently, we have built a model of the precise hydrogen bonding interactions between the third strand and, the duplex. Critical tests of the model are now possible; e.g., chemical probing (footprinting) of reactive groups in the third strand. In addition, we have searched for eukaryotic proteins other than recombinases that might either form or bind to three-stranded or triplex DNA. We have purified and characterized a protein from HeLa cells that binds to the TAT triplex. We used a gel shift assay to detect the protein. The protein is greatly enriched by passage through a triplex DNA-affinity column. Finally, in order to assess the role that homologous recombination might play in immunoglobulin heavy chain class switching, we have sequenced several human switch recombination sites. From an examination of these sequences, we have argued that they probably are not substrates for the homologous recombination machinery.